High-Performance Open-Source Archive
export_series().
Allows exporting loaded NASCAR series data to local CSV or Parquet
files, with automatic file type guessing based on file extension.ingesting-nascar-data vignette.
Documents direct access URLs, command-line download instructions, and
details on correctly parsing car numbers (preserving leading zeros) in R
and Python.pak::pak() for installing the package from CRAN
and GitHub.load_series()
now downloads fresh data from R2 once per session (in-memory cache
only).Bundled data removed. The
cup_series, xfinity_series, and
truck_series datasets are no longer shipped with the
package. All data is now served from Cloudflare R2. Use
load_series() to access data:
cup <- load_series("cup")
nxs <- load_series("nxs")
truck <- load_series("truck")arrow is now a required dependency
(moved from Suggests to Imports) since all data access goes through
parquet files on R2.
"xfinity" series renamed to
"nxs". The second-tier series has changed title
sponsors four times: Busch (1984-2007), Nationwide (2008-2014), Xfinity
(2015-2025), and O’Reilly Auto Parts (2026-present). The identifier
"nxs" is NASCAR’s own sponsor-neutral abbreviation, so it
will never go stale. Replace load_series("xfinity") with
load_series("nxs").
find_driver(), find_team(),
find_manufacturer() removed. The
get_*_info() functions already include fuzzy matching and
return actual data. Use get_driver_info("bell") instead of
find_driver("bell").
data("cup_series") no longer works.
Replace all data() calls with load_series().
The old lazy-loaded dataset names are gone.
load_series(): Downloads series
data from R2 and caches in memory. Subsequent calls within the same
session are instant. Use refresh = TRUE to force
re-download.
clear_cache(): New exported
function to reset the in-memory cache.
R2-canonical pipeline: The weekly GitHub Actions scraper now reads existing data from R2, appends new races, and uploads back to R2. No local rda files are generated or committed.
Migrated from httr to httr2 for HTTP
requests. The scraper now uses httr2::request() with
built-in retry logic (req_retry()).
Consolidated web scraping with httr2,
imap_dfr() indexing, explicit column type coercion,
placeholder detection, and empty table guards.
Data validation framework for schema, integrity, and value checks.
Code quality enforced with styler and
lintr (zero warnings).
All stop() calls in package code replaced with
rlang::abort().
Deprecating the weekly branch! Weekly race results will
be added to the main branch. CRAN-stable version is
available via CRAN:
install.packages("nascaR.data")
remotes::install_github("kyleGrealis/nascaR.data") # please do not use "@weekly"Interactive driver/team/manufacturer selection: When multiple matches are found, users can now select from a numbered list
Intelligent fuzzy matching: Dramatically improved search algorithm that handles typos, partial names, and word boundaries
find_driver("kyle") -> returns Kyle Busch, Kyle
Larson, Kyle Petty, etc.find_team("gibbs") -> finds Joe Gibbs Racingfind_driver("earnhart") -> correctly finds Earnhardt
family driversFlexible series input: All functions now accept both character strings AND data frames
get_driver_info("kyle", "cup") checkget_driver_info("kyle", "Cup Series") checkget_driver_info("kyle", cup_series) checkSmart string matching: Handles variations like “cup”, “Cup Series”, “xfinity”, “Xfinity Series” automatically
Consolidated codebase: Replaced three separate fuzzy matching files with one unified system
Priority-based matching: Exact matches > starts with > contains > word boundaries > fuzzy similarity
Non-interactive mode: Dashboard/script
developers can set interactive = FALSE to get list
returns
Removed dependency on problematic Levenshtein distance calculations
Eliminated interactive prompts that broke in non-interactive environments
Typo tolerance: Common misspellings now find correct matches
One-step workflow: Search and select in the same function call
Clear feedback: Better messaging when multiple options are available
None! All existing function calls continue to work as before
New interactive parameter defaults to
TRUE for better user experience
Fixed fuzzy matching returning irrelevant results
Resolved cases where obvious matches weren’t found due to strict string matching
Eliminated interactive readline prompts that failed in scripts and R Markdown
Added missing races. The Cup Series season finale was omitted for a number of years from 2002 to 2022. Thank you to Nick Triplett for the catching the mistake!
Seg Points has been removed. Instead,
S1 & S2 variables correspond to the
driver’s finishing position during each segment.
Updated missing track information (length, surface type) for 32 Cup races with varying years, mostly pre-2000s.
S1 & S2 documentation.dev branch will be updated with the most recent
racing results every Monday at 10AM during the race season (February
through November).
New get_*_info() functions provide summary
statistics on a by-race, season, or career format.
Fuzzy matching has been included to search across the database in the respective series
Error handling for placeholder race. The “DriverAverages.com” site will sometimes have one row for the upcoming race. This has routinely caused issues for the weekly scraping functions accidentally “recognizing” that as having been a completed race. The oversight has been addressed by removing that (essentially) blank row, decreasing the index, and continuing to scrape for new race data.
Fuzzy finding is used within the functions returning the driver, team, or manufacturer information. There is room for improvement: “Chris Bell” would find “Christopher Bell” instead of “Chris Miller” as it does now. Improvements would include a stronger focus on the driver or owner’s last name, but this will take some trial and error to really dial in.
Need mirroring services?
Contact our team at info@vpspulse.com.
Mirror powered by VPSpulse
Infrastructure sponsored by VPSPulse & Secure Payments by ArionPay.